Data selection

Previously, yellow and silver eel series have been analyzed by WGEEL through trend analysis (ICES 2021a). Here, we have chosen to concentrate on improving the trend analysis for the silver eel time series.

In the 2023 data call, 57 silver eel time series were available, located in 15 countries (Figure @ref(fig:mapsilverall)). The majority of these series are from Lithuania (9 series), Netherlands (7 series), United Kingdom (6 series) and France (6 series). Four time series (BI1S, BI4S, NSIS, PanS) are international trawling surveys. Data on these silver eel time series have been added to the WGEEL database in the past, but as these international series have no dedicated reporting country, their data have not been inserted into the database for recent years.

Map of available silver time series. Updated time series correspond to time series for which at least one value was provided for years the three last years

Summary of data availability in the silver eel time series, showing availability of series with at least 10 years of data; and the number of series that do or do not have available information on series quality ID, habitat type, sampling type, effort, gear type, restocking influence, and distance to sea.

Category	Available data	Missing data
Nb of series >=10 years	30	27
Nb of series with quality id	36	21
Nb of series with habitat data	57	0
Nb of series with sampling type	53	4
Nb of series with effort data	23	34
Nb of series with gear	56	1
Nb of series with restocking data	56	1
Nb of series with distance to sea	52	5

Summary of available silver eel series per country, including whether they were kept or dropped for the trend analysis. Habitat: C = coastal water, F = freshwater, MO = marine water (open sea), T = transitional water (according to WFD); gear: 226 = fyke nets, 227 =stow nets, 228 = barriers, fences, weirs, etc., 230 = traps, 234 = longlines, 242 = electric fishing, 245 = gear unknown; sampling type: 1 = commercial catch, 2 = commercial CPUE, 3 = scientific estimate, 4 = trapping all, 5 = trapping partial gear; quality id: 0 = series has ended and no future data will become available, 1 = good quality data; stocking: FALSE = no impacts of stocking, TRUE = impacts of stocking; and ecoregion: WM = Western Mediterranean Sea, BS = Baltic Sea, NS = Greater North Sea, IB = Bay of Biscay and the Iberian Coast, CS = Celtic Seas, AL = Aegean-Levantine Sea, CM = Ionian Sea and the Central Mediterranean Sea, AS = Adriatic Sea. For all, NA means not available (no information provided).

A summary of the type of data available in the available silver eel time series is given in Table @ref(tab:tablesummary) and Figure @ref(fig:figureseriessummary), and a detailed summary of all the series is presented in Annex x.x. From all available silver eel time series, 30 series have 10 or more years of available data. Only 23 of the 57 series have effort data submitted. Most silver eel series were collected in freshwaters via traps and fyke nets. In terms of sampling type, 14 series were from commercial catches, 5 series was reported as commercial CPUE, 10 were assigned as full trapping series, 18 as partial trapping series, and 26 were classified as scientific estimate, with 4 series missing this information. When it comes to the series quality ID, 38 series had a quality ID of 1, indicating good quality, 40 had a quality ID of 0, indicating that it had been previously decided that the series should not be used in an analysis (usually because there were not enough years of data), and 21 series were missing a series quality ID. For the next WGEEL meeting, these series quality IDs should be reconsidered, given the large amount of series without a series quality ID, and the fact that some series with a quality ID of 0 now have at least 10 years of data. Almost all series had information on the potential impacts of restocking, with 21 series classified as being influenced by restocking and 35 as not being influenced by restocking.

The decision was made to perform the trend analysis with only those silver eel time series that had at least 10 years of data available. After examining the 57 available silver eel time-series, only 28 were kept for the trend analysis (Table @ref(tab:tablekept). From the 29 time-series that were excluded, 2 were showing inconsistencies in data and 27 were excluded because they did not have 10 consecutive years of effort data (Table @ref(tab:tabledropped).

The 28 silver eel time series which were used in the analysis represented six ICES ecoregions (Figure @ref(fig:mapsilverkept)). The Western Mediterranean Sea and Aegean-Levantine Sea only had one time series. The most represented areas were the Greater North Sea with 8 time-series and the Baltic Sea with 7 time-series. 11 ecoregions (Adriatic Sea, Arctic Ocean, Azores, Barents Sea, Black Sea, Faroes, Greenland Sea, Icelandic Waters, Ionian Sea and the Central Mediterranean Sea, Norwegian Sea, Oceanic Northeast Atlantic) did not have silver eel time series which were used in the trend analysis. All coastal time series used in the analysis are located in the Baltic Sea ecoregion. Freshwater series are located across 4 different ecoregions (Baltic Sea, Bay of Biscay and the Iberian Coast, Celtic Seas, Greater North Sea) and the transitional series are located in 2 ecoregions (Aegean-Levantine Sea, Western Mediterranean Sea).

Available silver eel time series that were kept for the analysis. Listed is also whether the silver eel time-series data was first divided by provided effort data, before it was used in the analysis. A reason for dividing by effort is that the data in the series is effort-dependent. Reasons for not first dividing by effort data include for instance the provided data already being a CPUE index, or effort being constant (e.g. whole year full-river coverage).

Country	Series name	Years (n)	Series type	Divided by effort data	Comments
	BI1S	16	Index	No
	BI4S	20	Index	No
	NSIS	22	Index	No
	PanS	16	Index	No
NO	ImsaS	48	Constant effort	No
SE	NkaS	41	Index	No
SE	SosS	41	Index	No
DE	WarS	14	Model estimate	No
DK	RibS	22	Model estimate	No
NL	HVWS	10	Index	No
NL	NiWS	11	Index	No
IE	BurS	51	Constant effort	No
IE	KilS	23	Effort-dependent	Yes
GB	BaBS	17	Constant effort	No
GB	FowS	12	Constant effort	No
GB	GiBS	32	Constant effort	No
GB	LevS	21	Constant effort	No
GB	ShiS	20	Constant effort	No
FR	BreS	37	Effort-dependent	Yes
FR	FreS	27	Constant effort	No
FR	LoiS	34	Index	No
FR	SeNS	10	Constant effort	No
FR	SouS	10	Model estimate	No
ES	AlCS	68	Effort-dependent	No	Commercial catches, but assumed that all eel are caught.
ES	BidS	13	Constant effort	No
ES	NalS	11	Constant effort	No
ES	OriS	16	Index	No
GR	EamtS	11	Constant effort	No

Available silver eel time series that were dropped from the analysis, including the reason for being dropped.

Country	Series name	Years (n)	Drop reason	Comments
SE	KavlS	4	Number of years fewer than 10
FI	KotkS	6	Number of years fewer than 10
FI	VaakS	9	Number of years fewer than 10
LV	DaugS	4	Number of years fewer than 10
LV	LilS	3	Number of years fewer than 10
LT	AlauS	4	Number of years fewer than 10	Next year double check for available effort data on all similar Lithuanian series, and check if only silver eel are reported
LT	CurlS	5	Number of years fewer than 10	Next year double check for available effort data on all similar Lithuanian series, and check if only silver eel are reported
LT	KertS	4	Number of years fewer than 10	Next year double check for available effort data on all similar Lithuanian series, and check if only silver eel are reported
LT	KreS	1	Number of years fewer than 10	Next year double check for available effort data on all similar Lithuanian series, and check if only silver eel are reported
LT	LakS	2	Number of years fewer than 10	Next year double check for available effort data on all similar Lithuanian series, and check if only silver eel are reported
LT	RieS	1	Number of years fewer than 10	Next year double check for available effort data on all similar Lithuanian series, and check if only silver eel are reported
LT	RubS	2	Number of years fewer than 10	Next year double check for available effort data on all similar Lithuanian series, and check if only silver eel are reported
LT	SiesS	3	Number of years fewer than 10	Next year double check for available effort data on all similar Lithuanian series, and check if only silver eel are reported
LT	ZeiS	1	Number of years fewer than 10	Next year double check for available effort data on all similar Lithuanian series, and check if only silver eel are reported
NL	BRWS	8	Number of years fewer than 10
NL	DoijS	8	Number of years fewer than 10
NL	IjsS	8	Number of years fewer than 10
NL	NZKS	7	Number of years fewer than 10
NL	ZMaS	9	Number of years fewer than 10
GB	StrS	6	Number of years fewer than 10
FR	VilS	9	Number of years fewer than 10
PT	MinS	5	Number of years fewer than 10
PT	MonS	6	Number of years fewer than 10
IT	PobeS	1	Number of years fewer than 10
IT	PogoS	1	Number of years fewer than 10
IT	PolsS	1	Number of years fewer than 10
IT	TibeS	1	Number of years fewer than 10
GR	NorwS	10	Inconsistencies in data	Recent years reported as 0, but actually no information available
GR	WepeS	10	Inconsistencies in data	Recent years reported as 0, but actually no information available

Map of silver eel time-series used in the trend analysis, shown by ICES ecoregion and habitat.

Figure @ref(fig:figureseries) gives an overview of the reported CPUE data for each silver eel time series that was kept for the trend analysis. Among the kept series, the number of available CPUE data points was highest between 2010 and 2021, with the peak in 2013-2017 (Figure @ref(fig:figurenseries)). The majority of the kept time series did not have 2023 data ready at the time of writing this report.

CPUE data of those silver eel time series that were used in the trend analysis.

Number of available datapoints from silver eel CPUE time-series over time, shown only for those time-series that were used in the trend analysis.